Abstract

Species are fundamental units in evolutionary biology. However, defining them in taxonomically problematic groups requires integration of independent sources of information in order to develop robust hypotheses for taxonomic classification. Here, we propose an integrative framework for species delimitation in the Mediterranean species of the genus Cerithium (Caenogastropoda: Cerithiidae), whose shells show a wide variety of forms resulting in problematic morphological identification at the species level. Combined sequence data of two mitochondrial genes (COI+12S) for 55 individuals from the central Mediterranean Sea were used to test the species status of six identified morphotypes. Phylogenetic analyses, as well as DNA-based methods of species delimitation (automatic barcode gap discovery, species delimitation plugin and genealogical sorting index), support the species status of three morphotypes (C. vulgatum = MOTU-A, C. alucastrum = MOTU-B and C. protractum = MOTU-C), sometimes considered as ecotypes. Molecular operational taxonomic unit (MOTU)-D includes large individuals morphologically assigned to C. vulgatum (from Brindisi and Oristano) and C. repandum (from Tunisia), but all probably represent C. repandum which, if valid, would then not be endemic to its type locality in the Gulf of Gabès. All individuals identified as C. lividulum grouped in MOTU-E, except those from Tunisia which cluster in MOTU-F with some C. renovatum. Cerithium renovatum from Crete forms an endemic group (MOTU-G), suggesting a possible cryptic species. Our results show only partial concordance between traditional morphology and sequence data, indicating that the former is not always sufficient for recognizing species level taxa within Mediterranean Cerithium, although protoconch morphology is a key feature for distinguishing between species with superficially similar teleoconchs. Further analyses based on more comprehensive geographic sampling and more mitochondrial markers, and including a number of nuclear loci, are needed to clarify Mediterranean Cerithium diversity more fully.

INTRODUCTION

Delimiting species and defining their phylogenetic relationships are fundamental issues in evolutionary biology (Pavlinov, 2013). Historically, the taxonomy of a given organism is mainly based on morphological characters, whose variation is required for species delimitation. However, traditional morphology can fail to distinguish species in several cases, such as among closely related taxa, in very recent radiations, or in groups that exhibit overlapping intra- and interspecific variation (i.e. Wiens, 2007; Schlick-Steiner et al., 2010). As a result, morphotypes do not always correspond to actual species, but can reflect varieties or ecophenotypes within a single species or admixtures of more than one species. This is especially true for several highly diverse invertebrate groups, for which it is difficult to estimate species richness by employing morphological criteria alone. In these cases, DNA sequence data have supplied an independent source of characters useful for identifying species (i.e. Sites & Marshall, 2003; Pons et al., 2006). Thus, an integrative taxonomic approach, including both morphological and molecular data, has become the contemporary way to explore species richness, particularly in the case of taxonomically problematic groups (e.g. Dayrat, 2005; Fujita et al., 2012). This approach is also well in agreement with the phylogenetic species concept (PSC) or the more general lineage concept of species (de Queiroz, 1998, 2005a, b, 2007), both of which identify species as a diverging lineage and overcome the problem of the different species concepts (Mallet, 2001).

In the marine environment, a particularly striking example illustrating confusing species-level taxonomy is represented by the family Cerithiidae, a group of primarily marine caenogastropods including hundreds of species distributed worldwide, from temperate to tropical seas. Most of them live in shallow waters on both soft and hard substrates, especially in seagrass beds, lagoons and brackish-water systems. They are well known for displaying an impressive intraspecific variability in shell morphology, causing great taxonomic confusion. As a consequence, species delimitation within the current morphology-based system is often problematic, especially when authors give different importance to morphological features. The disjunct distributions of some species can create further taxonomic complexity, because there is no standardized method to establish whether allopatric populations are conspecific. Systematic revisions of the largest genera were completed by Houbrick (1974, 1976, 1985, 1992), but focused only on Western Atlantic and Indo-Pacific species. Several other studies considered the Mediterranean Cerithium fauna and the most conservative view (Bucquoy, Dautzemberg & Dolfus, 1884) recognized only two species, the ‘small’ littoral C. rupestre Risso, 1826 and the ‘large’ subtidal C. vulgatum Bruguière, 1792. Several revisions at the beginning of the 20th century (e.g. Monterosato, 1910; Nordsieck, 1974) only added to the confusion, recognizing from two to 52 taxa belonging to different subgenera. According to a recent checklist (Cecalupo, 2006), four native species currently inhabit the Mediterranean Sea: C. alucastrumBrocchi, 1814, C. lividulum Risso, 1826, C. renovatumMonterosato, 1884 and C. vulgatum.Giannuzzi-Savelli et al. (1997) and Oliverio et al. (2008) also considered C. haustellum Monterosato in Crema, 1803 and C. protractum Ant. Bivona in And. Bivona, 1838 to be valid, but these were listed as possible synonyms by Gofas (2015). A recent study using isozyme electrophoresis by Boisselier-Dubayle & Gofas (1999) supported the recognition of two small species, C. lividulum and C. renovatum (as C. rupestre; Gofas, Garilli & Boisselier-Dubayle, 2004), both with nonplanktotrophic larvae, and supported the division of C. vulgatum into two morphologically similar large species: true C. vulgatum from open-sea habitats with planktotrophic development and a ‘lagoon’ species with nonplanktotrophic development.

In the present study, an integrative taxonomic approach was implemented to assess the diversity among native Mediterranean Cerithium and to provide a robust framework for addressing the following questions: (1) Do genetics and morphology in the native Mediterranean Cerithium vary in a similar manner? (2) Are the current morphological characters used in the taxonomy of Cerithium valid for species identification? (3) How many native Cerithium species currently inhabit the Mediterranean Sea and what are the relationships among them?

MATERIAL AND METHODS

Sampling and morphological identification

Specimens of the genus Cerithium were collected at 16 central Mediterranean localities, from the north Adriatic Sea to North Africa and from Sardinia to Crete, spanning c. 1,275 and 1,490 km respectively (Fig. 1). Upon collection by hand, specimens were fixed in absolute ethanol and kept at room temperature. All snails were then assigned to known taxa according to their morphology, using a combination of shell size, teleoconch sculpture, protoconch morphology, headfoot colouration and radular morphology (see Table 1). Typically, only a subset of these characters was used; the first whorls of the shell are often eroded, preventing the analysis of the protoconch, and ethanol preservation makes observations on colour of the animals unreliable. Our identifications are therefore mainly based on teleoconch sculpture. All studied material is housed in the Zoological Museum of the University of Bologna, Italy.

Table 1.

Morphological characters traditionally used to discriminate nominal species of Cerithium.

 Size Teleoconch sculpture Protoconch Protoconch sculpture Radula Habitat Egg masses 
C. alucastrum >70 mm height/body whorl c.1.5 Axial folds, no spiny knobs (Brocchi, 1814; Monterosato, 1910Planktotrophic (Montserrat, 1994Protoconch 1 smooth (Montserrat, 1994); protoconch 2 unknown Unknown Circalittoral Gelatinous, irregularly coiled string, several eggs across (Montserrat, 1994
C. vulgatum c. 40–60 mm height/body whorl <1.5 Spiny knobs (Cecalupo et al., 2008Planktotrophic (Cecalupo et al., 2008Protoconch 1 almost smooth (Cecalupo et al., 2008); protoconch 2 with axial subsutural lines and spiral rows of dots (Cecalupo et al., 20083 ectocones on rachidian tooth? (Cecalupo et al., 2008Infralittoral, open sea, rocks, Posidonia beds, algal bottoms (Cecalupo et al., 2008Unknown 
C. protractum c. 40 mm height/body whorl c. 1.5 Spiny knobs (Gofas et al., 2011Unknown Unknown Unknown Deep infralittoral (Gofas et al., 2011Unknown 
C. repandum c. 40 mm height/body whorl <1.5 Spiny knobs (Cecalupo et al., 2008Nonplanktotrophic (Cecalupo et al., 2008Almost smooth (Cecalupo et al., 20082 ectocones on rachidian tooth? (Cecalupo et al., 2008Upper infralittoral, lagoons, fine sand or mud (Cecalupo et al., 2008Unknown 
C. lividulum c. 20 mm Axial folds (Gofas et al., 2004; Garilli & Galletti, 2006Nonplanktotrophic (Gofas et al., 2004; Garilli & Galletti, 2006Spiral rows of dots, whorls regularly rounded (Garilli & Galletti, 2006Unknown Mesolittoral and upper infralittoral (Gofas et al., 2011Massive, gelatinous, more or less cylindrical 4–5 eggs across (Palombi, 1939; Barash & Zenziper, 1980; Garilli & Galletti, 2006
C. renovatum c. 20 mm Knobs (Garilli & Galletti, 2006Nonplanktotrophic (Gofas et al., 2004; Garilli & Galletti, 2006Spiral rows of microtubercles, whorls angulate (Garilli & Galletti, 2006Unknown Upper infralittoral (Gofas et al., 2011Gelatinous string, 1 egg across (Garilli & Galletti, 2006
 Size Teleoconch sculpture Protoconch Protoconch sculpture Radula Habitat Egg masses 
C. alucastrum >70 mm height/body whorl c.1.5 Axial folds, no spiny knobs (Brocchi, 1814; Monterosato, 1910Planktotrophic (Montserrat, 1994Protoconch 1 smooth (Montserrat, 1994); protoconch 2 unknown Unknown Circalittoral Gelatinous, irregularly coiled string, several eggs across (Montserrat, 1994
C. vulgatum c. 40–60 mm height/body whorl <1.5 Spiny knobs (Cecalupo et al., 2008Planktotrophic (Cecalupo et al., 2008Protoconch 1 almost smooth (Cecalupo et al., 2008); protoconch 2 with axial subsutural lines and spiral rows of dots (Cecalupo et al., 20083 ectocones on rachidian tooth? (Cecalupo et al., 2008Infralittoral, open sea, rocks, Posidonia beds, algal bottoms (Cecalupo et al., 2008Unknown 
C. protractum c. 40 mm height/body whorl c. 1.5 Spiny knobs (Gofas et al., 2011Unknown Unknown Unknown Deep infralittoral (Gofas et al., 2011Unknown 
C. repandum c. 40 mm height/body whorl <1.5 Spiny knobs (Cecalupo et al., 2008Nonplanktotrophic (Cecalupo et al., 2008Almost smooth (Cecalupo et al., 20082 ectocones on rachidian tooth? (Cecalupo et al., 2008Upper infralittoral, lagoons, fine sand or mud (Cecalupo et al., 2008Unknown 
C. lividulum c. 20 mm Axial folds (Gofas et al., 2004; Garilli & Galletti, 2006Nonplanktotrophic (Gofas et al., 2004; Garilli & Galletti, 2006Spiral rows of dots, whorls regularly rounded (Garilli & Galletti, 2006Unknown Mesolittoral and upper infralittoral (Gofas et al., 2011Massive, gelatinous, more or less cylindrical 4–5 eggs across (Palombi, 1939; Barash & Zenziper, 1980; Garilli & Galletti, 2006
C. renovatum c. 20 mm Knobs (Garilli & Galletti, 2006Nonplanktotrophic (Gofas et al., 2004; Garilli & Galletti, 2006Spiral rows of microtubercles, whorls angulate (Garilli & Galletti, 2006Unknown Upper infralittoral (Gofas et al., 2011Gelatinous string, 1 egg across (Garilli & Galletti, 2006

See Figure 3 for protoconchs of C. vulgatum, C. repandum, C. lividulum and C. renovatum.

Figure 1.

Map of 16 sampled localities for Cerithium specimens in the central Mediterranean Sea. Labels: a, Chioggia (Italy); b, Trieste (Italy); c, Spalato (Croatia); d, Pag (Croatia); e, Pasman (Croatia); f, Monopoli (Apulia, Italy); g, Brindisi (Apulia, Italy); h, Gallipoli (Apulia, Italy); i, Torre Colimena (Apulia, Italy); l, Crete (Greece); m, Marzamemi (Sicily, Italy); n, Lampedusa (Italy); o, Djerba (Tunisia); p, Kerkennah (Tunisia); q, Oristano (Sardinia, Italy); r, Elba (Italy).

Figure 1.

Map of 16 sampled localities for Cerithium specimens in the central Mediterranean Sea. Labels: a, Chioggia (Italy); b, Trieste (Italy); c, Spalato (Croatia); d, Pag (Croatia); e, Pasman (Croatia); f, Monopoli (Apulia, Italy); g, Brindisi (Apulia, Italy); h, Gallipoli (Apulia, Italy); i, Torre Colimena (Apulia, Italy); l, Crete (Greece); m, Marzamemi (Sicily, Italy); n, Lampedusa (Italy); o, Djerba (Tunisia); p, Kerkennah (Tunisia); q, Oristano (Sardinia, Italy); r, Elba (Italy).

Molecular protocols and sequence alignment

For each identified morphospecies, 3–18 individuals were photographed and then cracked in order to access the soft tissue for molecular analyses. A total of 55 individuals were used. Total genomic DNA was isolated from a small piece of foot muscle using the CTAB protocol for molluscs (Winnepenninckx, Backeljau & de Wachter, 1993) and stored at −20 °C prior to amplification. Two DNA fragments from the mitochondrial genome (COI and 12SrDNA) were amplified with the primers listed in Table 2. For COI, two forward and two reverse primers were used in four different combinations (LCO1490/COIbr, LCO1490/COI6, COIbf/COIbr and COIbf/COI6) and each pair was optimized over a range of annealing temperatures from 40 to 51 °C. The COIbf/COIbr combination was the most successful. All PCRs were carried out in a total volume of 25 µl with 2.5 µl dNTPs (2 mM each; Promega), 2.5 µl 10× load buffer-MgCl2 (Fermentas), 2.5 µl of each primer (2 µM each; Invitrogen), 0.25 µl Taq (5 U/µl; Fermentas), 3 µl bovine serum albumin (BioLabs), 20 ng template DNA and purified water. PCR cycling conditions were as follows: for COI, initial denaturation at 94 °C for 3 min, 40 cycles of denaturation at 94 °C for 1 min, annealing at 48 °C (LCO1490/COIbr), 43 °C (LCO1490/COI6), 51 °C (COIbf/COIbr) and 42 °C (COIbf/COI6), and a final extension at 72 °C for 7 min; for 12 S, initial denaturation at 95 °C for 5 min, 40 cycles of denaturation at 95 °C for 45 s, annealing at 51 °C for 45 s and a final extension at 72 °C for 5 min. All amplifications were run on a Gene Amp® PCR System 2700 (Applied Biosystems) and included positive and negative controls. PCR products were visualized on 1.5% agarose/TAE1x gel, stained with SYBR Safe 10× concentrate in DMSO (Invitrogen) and quantified using a ND-1,000 spectrophotometer. Unincorporated dNTPs, primers and other impurities were removed using the Wizard® SV Gel and PCR Clean-Up System (Promega), following the manufacturer's instructions. Purified amplicons were then sent to Macrogen Corporation (Rockville, USA) for bidirectional sequencing using the same primers mentioned above (5 µM each). Chromatograms were verified and manually corrected in Chromas Lite v. 2.1.1 (http://www.technelysium.com.au/chromas_lite.html) and consensus sequences assembled in BioEdit v. 7.1.11 (Hall, 1999). Alignments for the three genes were generated using ClustalW (Thomspon, Higgins & Gibson, 1994) in MEGA5 (Tamura et al., 2011) with default parameters. COI sequences were translated into amino acids using MEGA5 to check for frame shifts and stop codons. Number of variable sites, number of parsimony informative sites, number of gap positions or missing data, number of haplotypes, haplotype diversity and nucleotide diversity were calculated in DnaSP v. 5.10 (Librado & Rozas, 2009). Levels of saturation for each gene and for COI codon position were assessed using the substitution saturation test of Xia et al. (2003), implemented in DAMBE v. 5.3.74 (Xia & Xie, 2001). To test for statistical congruence between the two genes, the incongruence length difference (ILD) test was performed with PAUP* v. 4.0b (as the partition homogeneity test; Swofford, 2002), using 1,000 replicates and the standard heuristic search options (P > 0.05: allowed combined dataset).

Table 2.

Gene fragments and primers used in this study.

Gene fragment Primer Primer sequence 5′-3′ Forward (F) or reverse (R) Authors 
COI LCO1490 GGTCAACAAATCATAAAGATATTGG (25-mer) Folmer et al. (1994) 
COIbf GGGGCTCCTGATATAGCTTTTCC (23-mer) Miura et al. (2006) 
COIbr TAATATAGAAGTGTGCTTTAGT (22-mer) Miura et al. (2006) 
COI6 GGRTARTCNSWRTANCGNCGNGG (23-mer) Shimayama et al. (1990) 
12S 12SaL AAACTGGGATTAGATACCCCACTAT (25-mer) Kocher et al. (1989) 
12SaH GAGGGTGACGGGCGGTGTGT (20-mer) Kocher et al. (1989) 
Gene fragment Primer Primer sequence 5′-3′ Forward (F) or reverse (R) Authors 
COI LCO1490 GGTCAACAAATCATAAAGATATTGG (25-mer) Folmer et al. (1994) 
COIbf GGGGCTCCTGATATAGCTTTTCC (23-mer) Miura et al. (2006) 
COIbr TAATATAGAAGTGTGCTTTAGT (22-mer) Miura et al. (2006) 
COI6 GGRTARTCNSWRTANCGNCGNGG (23-mer) Shimayama et al. (1990) 
12S 12SaL AAACTGGGATTAGATACCCCACTAT (25-mer) Kocher et al. (1989) 
12SaH GAGGGTGACGGGCGGTGTGT (20-mer) Kocher et al. (1989) 

Phylogenetic analyses and species delimitation

Identical sequences were collapsed in the resulting trees, which were inferred for each gene separately and for the COI-12S concatenated dataset (cmtDNA). Phylogenetic analyses were performed to define molecular operational taxonomic units (MOTUs; Blaxter, 2004), i.e. well supported clades representing potential species under the PSC. The Bayesian information criterion (BIC; Schwarz, 1978) and the corrected Akaike information criterion (AIC; Sugiura, 1978; Hurvich & Tsai, 1989) were employed in MEGA5 to determine the best-fit model of evolution, selected for each gene separately, and for each codon position. Phylogenetic trees were inferred by Neighbour Joining (NJ), Maximum Parsimony (MP), Maximum likelihood (ML) and Bayesian Inference (BI) methods. Trees were rooted with specimens of C. scabridum (this species was sampled in Tunisia, but is endemic to the NW Indian Ocean and is introduced in the Mediterranean). NJ, MP and ML analyses were conducted in MEGA5 using default parameters, with complete deletion of missing data and alignment gaps (indels). Nodal support was assessed using nonparametric bootstrapping with 1,000 replicates, and only considered significant above 80% (Kawahara et al., 2011). BI searches were executed in MrBayes v. 3.2 (Ronquist et al., 2012). The Markov-chain Monte Carlo analysis was run for 3,000,000 generations, saving the current tree file every 100 generations. Four partitions (corresponding to 12S and each COI codon position) were set for the cmtDNA dataset, with the appropriate nucleotide substitution model applied to each partition. Alignment gaps were treated as missing. Topologies prior to –ln likelihood stabilization were discarded as burn-in and a majority rule consensus tree, as well as the Bayesian posterior probability (BPP) values, were computed from the remaining trees. In all cases, stationarity was reached within the 15,000th generation. Only BPP values equal to or above 95% were considered significant (Erixon et al., 2003). The obtained BI trees were drawn and edited using Dendroscope3 (Huson & Scornavacca, 2012).

Four methods were used to explore species boundaries with the cmtDNA dataset, to assess the taxonomic distinctiveness of species inferred from phylogenetic analysis: (1) DNA barcoding gap (Hebert et al., 2003); (2) species delimitation plugin (SDP; Masters, Fan & Ross, 2011); (3) genealogical sorting index (GSI; Cummings, Neel & Shaw, 2008) and (4) generalized mixed Yule coalescent (GMYC; Pons et al., 2006; Fontaneto et al., 2007). By analysing the distribution of pairwise distances among aligned sequences, the DNA barcoding gap method seeks to detect a break, the ‘barcode gap’, between intraspecific and interspecific values to assign the sequences to putative species (‘preliminary species hypotheses’, PSH), with two DNA sequences considered members of different species if the pairwise distance exceeds a certain threshold value. We applied the automatic barcode gap discovery (ABGD) method (Puillandre et al., 2012) which, rather than using a single a priori threshold, evaluates a range of thresholds obtained from the data themselves. The ABGD method was implemented using the Web interface at http://wwwabi.snv.jussieu.fr/public/abgd (last accessed 29 October 2015) on the cmtDNA dataset and on the COI dataset alone, using the K2P model to calculate pairwise distances (K80 option with TS/TV = 2.0), 20 recursive steps, X (relative gap width) = 0.5 and the remaining parameters set to default values (Pmin = 0.001, Pmax = 0.1, Nb bins = 20).

The SDP in Geneious (http://geneious.com, last accessed 1 November 2015) evaluates the probability that an observed monophyetic clade could have occurred by chance alone as a result of the coalescent process. It also assesses intra- and interspecific genetic distances to infer the probability that members of a putative species could be identified correctly. Rosenberg's reciprocal monophyly PAB (Rosenberg, 2007) and Rodrigo's randomly distinct P(RD) (Rodrigo et al., 2008) were estimated for each clade of the cmtDNA BI tree, with a PAB value smaller than 10−5 considered significant and a P(RD) value smaller than 0.05 defining a distinct species.

The GSI provides a measure of the relative degree of exclusive ancestry of a given group on a phylogenetic tree, where the maximum value of one indicates monophyly and the minimum of zero indicates dispersal over the entire tree. Statistical significance is assessed with a permutation test that generates trees with randomly rearranged individuals. The frequency of GSI values for a group of individuals in the permuted trees that are equal to or greater than the GSI in the original tree provides the P-value for rejection of the null hypothesis that the group is of mixed ancestry. We ran the second BI analysis on the cmtDNA dataset after removing the two divergent C. scabridum haplotypes (unrooted tree) and imported the tree into the website http://www.genealogicalsorting.org. Individuals were assigned to groups according to the MOTUs (Blaxter, 2004) identified in the phylogenetic analyses and the trees tips were permuted 10,000 times to reconstruct the null distribution.

The GMYC method infers the transition from coalescent population processes to phylogenetic Yule processes on an ultrametric tree with branch lengths scaled to time using a likelihood-based analysis to detect the threshold value between within-species and inter-species branching rates. The likelihood of the GMYC model is then compared to a null model assuming that all sequences belong to a single species. After removal of the outgroup (C. scabridum), the cmtDNA tree was imported into the R environment (http://www.r-project.org, last accessed 10 December 2015) using the ‘read.nexus’ command in the package ‘ape’. The tree was converted into an ultrametric and fully dichotomous tree using the commands ‘chronopl’ and ‘multi2di’, respectively, and subjected to the GMYC analysis as implemented in the SPLITS program (v. 2.10) in R (available at http://r-forge.r-project.org/projects/splits/, last accessed 20 June 2014). Both the single-threshold (single point of transition estimated; Pons et al., 2006) and multiple-threshold (point of transition allowed to vary among lineages; Monaghan et al., 2009) models were applied to the data.

RESULTS

Morphological identification

Excluding the outgroup Cerithium scabridum, the 55 analysed Cerithium specimens from the central Mediterranean Sea were classified into six morphotypes (Table 3, Fig. 2). Our C. alucastrum, C. protractum and C. vulgatum from the Adriatic Sea conformed well to the classical descriptions and figures of the species and were easily identified. Two large specimens from Sardinia (O1) and from Apulia (24) were assigned doubtfully to C. vulgatum as they showed less typical teleoconch characters compared with individuals from the Adriatic Sea. Based on their similar morphology, all other large specimens from Tunisia and Lampedusa were tentatively tagged as C. repandum, a species considered endemic to the Gulf of Gabès, and were collected in its typical habitat of very shallow water on sandy substrate. In addition, very young individuals with paucispiral protoconchs typical of C. repandum were found in the sediment collected at the sampling site. Finally, the morphological separation of the two small species C. lividulum and C. renovatum was based on the presence of knobs on the shell surface, more developed on the penultimate whorl of C. renovatum. However this identification could have been partly incorrect. In this case, the lack of the protoconch was particularly restrictive, as this seems to be the only character capable of reliably distinguishing between the two (Garilli & Galletti, 2006; Fig. 3).

Table 3.

Morphospecies of Cerithium, collecting sites, individuals sampled at each site and scored haplotypes.

Morphospecies Collecting site Country Individual Haplotypes
 
COI 12S cmtDNA 
C. vulgatum Pag Croatia CV1 a1 b1 cmt1 
   CV3 a2 b2 cmt2 
   CV6 a2 b2 cmt2 
 Pasman  A3 a3 b3 cmt3 
   A4 a4 b4 cmt4 
 Spalato  B1 a5 b1 cmt5 
   B2 a6 b5 cmt6 
   B3 a7 b5 cmt7 
   C1 a8 b5 cmt8 
   C2 a9 b5 cmt9 
   C3 a10 b5 cmt10 
 Trieste Italy CVT1 a11 b5 cmt11 
   CVT2 a2 b2 cmt2 
   CVT3 a2 b5 cmt12 
 Chioggia  CVV3 a12 b2 cmt13 
   CVV5 a13 b5 cmt14 
 Oristano, Sardinia  O1 a14 b6 cmt15 
 Brindisi, Apulia  24 a15 b6 cmt16 
C. protractum Pag Croatia CH1 a16 b7 cmt17 
   CH3 a17 b8 cmt18 
   CH5 a18 b8 cmt19 
   CH6 a19 b9 cmt20 
C. alucastrum Chioggia Italy CA1 a20 b10 cmt21 
   CA2 a21 b11 cmt22 
   CA3 a22 b10 cmt23 
C. repandum Lampedusa  CVL1 a23 b12 cmt24 
   CVL2 a24 b13 cmt25 
 Djerba Tunisia CR3 a25 b6 cmt26 
   CR4 a26 b6 cmt27 
 Kerkennah  C1 a27 b14 cmt28 
   C2 a28 b6 cmt29 
C. renovatum Elba Italy CREE1 a29 b15 cmt30 
   CREE2 a30 b16 cmt31 
 Brindisi, Apulia  23 a31 b17 cmt32 
   CRE2 a32 b18 cmt33 
   CRE3 a33 b19 cmt34 
   CRE14 a33 b19 cmt34 
   CRE15 a33 b19 cmt34 
 Crete Greece CRE16 a34 b20 cmt35 
   CRE26 a35 b21 cmt36 
   CRE27 a36 b22 cmt37 
C. lividulum Pag Croatia CLP1 a37 b23 cmt38 
   CLP2 a37 b23 cmt38 
   CLP3 a37 b23 cmt38 
   CLP4 a37 b23 cmt38 
 Monopoli, Apulia Italy CL7 a37 b23 cmt38 
   CL8 a37 b23 cmt38 
   CL9 a37 b23 cmt38 
 Marzamenni, Sicily  CL1 a38 b23 cmt39 
   CL3 a39 b24 cmt40 
   CL5 a40 b24 cmt41 
   CL6 a40 b24 cmt41 
 Djerba Tunisia CLD3 a41 b18 cmt42 
C. scabridum   CS6 a42 b25 cmt43 
   CS11 a43 b26 cmt44 
Morphospecies Collecting site Country Individual Haplotypes
 
COI 12S cmtDNA 
C. vulgatum Pag Croatia CV1 a1 b1 cmt1 
   CV3 a2 b2 cmt2 
   CV6 a2 b2 cmt2 
 Pasman  A3 a3 b3 cmt3 
   A4 a4 b4 cmt4 
 Spalato  B1 a5 b1 cmt5 
   B2 a6 b5 cmt6 
   B3 a7 b5 cmt7 
   C1 a8 b5 cmt8 
   C2 a9 b5 cmt9 
   C3 a10 b5 cmt10 
 Trieste Italy CVT1 a11 b5 cmt11 
   CVT2 a2 b2 cmt2 
   CVT3 a2 b5 cmt12 
 Chioggia  CVV3 a12 b2 cmt13 
   CVV5 a13 b5 cmt14 
 Oristano, Sardinia  O1 a14 b6 cmt15 
 Brindisi, Apulia  24 a15 b6 cmt16 
C. protractum Pag Croatia CH1 a16 b7 cmt17 
   CH3 a17 b8 cmt18 
   CH5 a18 b8 cmt19 
   CH6 a19 b9 cmt20 
C. alucastrum Chioggia Italy CA1 a20 b10 cmt21 
   CA2 a21 b11 cmt22 
   CA3 a22 b10 cmt23 
C. repandum Lampedusa  CVL1 a23 b12 cmt24 
   CVL2 a24 b13 cmt25 
 Djerba Tunisia CR3 a25 b6 cmt26 
   CR4 a26 b6 cmt27 
 Kerkennah  C1 a27 b14 cmt28 
   C2 a28 b6 cmt29 
C. renovatum Elba Italy CREE1 a29 b15 cmt30 
   CREE2 a30 b16 cmt31 
 Brindisi, Apulia  23 a31 b17 cmt32 
   CRE2 a32 b18 cmt33 
   CRE3 a33 b19 cmt34 
   CRE14 a33 b19 cmt34 
   CRE15 a33 b19 cmt34 
 Crete Greece CRE16 a34 b20 cmt35 
   CRE26 a35 b21 cmt36 
   CRE27 a36 b22 cmt37 
C. lividulum Pag Croatia CLP1 a37 b23 cmt38 
   CLP2 a37 b23 cmt38 
   CLP3 a37 b23 cmt38 
   CLP4 a37 b23 cmt38 
 Monopoli, Apulia Italy CL7 a37 b23 cmt38 
   CL8 a37 b23 cmt38 
   CL9 a37 b23 cmt38 
 Marzamenni, Sicily  CL1 a38 b23 cmt39 
   CL3 a39 b24 cmt40 
   CL5 a40 b24 cmt41 
   CL6 a40 b24 cmt41 
 Djerba Tunisia CLD3 a41 b18 cmt42 
C. scabridum   CS6 a42 b25 cmt43 
   CS11 a43 b26 cmt44 
Figure 2.

Main morphotypes of the native Mediterranean Cerithium specimens included in this study. All illustrated specimens were used for DNA extraction and are representatives of both identified morphospecies and MOTUs recovered by genetic analyses. See Table 3 for scored combined mitochondrial (cmt) DNA haplotypes. A.C. alucastrum (haplotype cmt23) B.C. protractum (cmt19). C–E.C. vulgatum (cmt1, cmt2 and cmt14, respectively). F, G.C. vulgatum (according to morphological identification, but C. repandum according to molecular evidence; cmt15 and cmt16, respectively). H, I.C. repandum (cmt24 and cmt27, respectively). J–L.C. lividulum (cmt37, cmt38 and cmt39, respectively). M.C. lividulum (according to morphological identification, but C. renovatum according to molecular evidence; cmt42). N.C. renovatum (cmt31). O.C. renovatum (according to morphological identification, but probably a new species according to molecular evidence; cmt35). P.C. renovatum (cmt34).

Figure 2.

Main morphotypes of the native Mediterranean Cerithium specimens included in this study. All illustrated specimens were used for DNA extraction and are representatives of both identified morphospecies and MOTUs recovered by genetic analyses. See Table 3 for scored combined mitochondrial (cmt) DNA haplotypes. A.C. alucastrum (haplotype cmt23) B.C. protractum (cmt19). C–E.C. vulgatum (cmt1, cmt2 and cmt14, respectively). F, G.C. vulgatum (according to morphological identification, but C. repandum according to molecular evidence; cmt15 and cmt16, respectively). H, I.C. repandum (cmt24 and cmt27, respectively). J–L.C. lividulum (cmt37, cmt38 and cmt39, respectively). M.C. lividulum (according to morphological identification, but C. renovatum according to molecular evidence; cmt42). N.C. renovatum (cmt31). O.C. renovatum (according to morphological identification, but probably a new species according to molecular evidence; cmt35). P.C. renovatum (cmt34).

Figure 3.

Protoconchs of four Mediterranean Cerithium species. A.C. vulgatum, planktotrophic. B.C. repandum, nonplanktotrophic. C.C. renovatum, nonplanktotrophic. D.C. lividulum, nonplanktotrophic.

Figure 3.

Protoconchs of four Mediterranean Cerithium species. A.C. vulgatum, planktotrophic. B.C. repandum, nonplanktotrophic. C.C. renovatum, nonplanktotrophic. D.C. lividulum, nonplanktotrophic.

Phylogenetic analyses

The basic statistics of the analysed alignments are summarized in Table 4. No stop codons were observed in the COI sequences. The saturation analysis showed insignificant levels of saturation for both the genes, COI: ISS (0.132) < ISS.C (0.818), P = 0.000, and 12S: ISS (0.064) < ISS.C (0.682), P = 0.000, even when third codon positions were analysed independently: ISS (0.292) < ISS.C (0.782), P = 0.000. Models of DNA substitution and tree likelihood scores are given in Table 4. The ILD test was not significant (P = 0.69), indicating that the two mitochondrial genes could be analysed in a combined dataset. Trees based on individual mitochondrial gene fragments (Supplementary Material Figs S1–S8) and on the cmtDNA dataset showed very similar topologies, with only a few differences, so we focused only on the cmtDNA phylogeny. NJ, MP, ML and BI cmtDNA trees revealed very similar patterns, showing congruence among them (Shimodaira-Hasegawa SH test not significant, P = 0.283). Therefore, only the BI cmtDNA tree is shown, along with the BPP and the NJ, MP and ML bootstrap values for the main nodes (Fig. 4). The tree clustered the 44 combined haplotypes into seven highly divergent and strongly supported lineages, only some of which were congruent with the morphological analysis: one including the northern populations (Adriatic Sea) of C. vulgatum (MOTU-A); two corresponding to C. alucastrum and C. protractum (MOTU-B and MOTU-C); one representing C. repandum + the western (cmt15 from Sardinia) and southern (cmt16 from Apulia) individuals of C. vulgatum (MOTU-D); one including C. lividulum from Croatia, Apulia and Sicily (MOTU-E); one corresponding to C. renovatum + the most southern (Tunisian) populations of C. lividulum (MOTU-F) and, finally, one including C. renovatum from Crete (MOTU-G). Identified morphotypes of C. alucastrum and C. protractum were supported as monophyletic, while those of C. vulgatum, C. lividulum and C. renovatum revealed polyphyletic patterns. All individuals identified as C. repandum clustered together, but they did not form a monophyletic group. The relationships among the seven clades were highly supported, with sister relationships supported between the northern lineage of C. vulgatum (MOTU-A) and C. alucastrum (MOTU-B), as well as between C. protractum (MOTU-C) and [MOTU-A+MOTU-B] (=GROUP1; Fig. 4). In addition, MOTU-D and MOTU-E appeared as sister groups (=GROUP2; Fig. 4), as well as MOTU-F and MOTU-G (=GROUP3; Fig. 4). There was also strong support for the monophyly of [GROUP1+GROUP2].

Table 4.

Sequence information for COI and 12S gene fragments (computed without the outgroup).

Gene fragment N Length PS PI G/M n h π Model Parameters BIC AICc -ln L 
COI 53 944 (302) (286) (1) (43) (0.979 ± 0.011) (0.106 ± 0.005) HKY + G 108 8807.1 7855.1 3819.3 
COI position1 – – – – – – – – T92 + I 106 5126.4 4309.5 2048.0 
COI position2 – – – – – – – – K2 + G 105 3065.1 2255.5 1022.1 
COI position3 – – – – – – – – HKY 87 1814.9 1162.6 493.72 
12S 53 372 (72) (65) (5) (26) (0.945 ± 0.015) (0.046 ± 0.003) T92 + G 52 2625.8 2253.6 1074.5 
cmtDNA 53 1316 (373) (349) (6) (44) (0.981 ± 0.011) (0.089 ± 0.004) HKY + G 90 11829 11022 5421.1 
Gene fragment N Length PS PI G/M n h π Model Parameters BIC AICc -ln L 
COI 53 944 (302) (286) (1) (43) (0.979 ± 0.011) (0.106 ± 0.005) HKY + G 108 8807.1 7855.1 3819.3 
COI position1 – – – – – – – – T92 + I 106 5126.4 4309.5 2048.0 
COI position2 – – – – – – – – K2 + G 105 3065.1 2255.5 1022.1 
COI position3 – – – – – – – – HKY 87 1814.9 1162.6 493.72 
12S 53 372 (72) (65) (5) (26) (0.945 ± 0.015) (0.046 ± 0.003) T92 + G 52 2625.8 2253.6 1074.5 
cmtDNA 53 1316 (373) (349) (6) (44) (0.981 ± 0.011) (0.089 ± 0.004) HKY + G 90 11829 11022 5421.1 

Abbreviations: N, sample size; length, fragment length (bp); PS, number of variable sites; PI, number of parsimony informative sites; G/M, number of gap positions or missing data; n, haplotype number; h ± SD, haplotype diversity ± standard deviation; π ± SD, nucleotide diversity ± standard deviation; model, evolution models selected in MEGA5; for each model: parameters, number of model parameters; BIC, Bayesian information criterion; AICc, corrected Akaike information criterion; –ln L, maximum likelihood value.

Figure 4.

Bayesian phylogenetic tree for the native Mediterranean Cerithium specimens based on the combined COI and 12S mitochondrial gene (cmtDNA) dataset and rooted with C. scabridum. Names of terminal taxa correspond to the scored haplotypes (listed in Table 3). Morphotypes doubtfully (?) included indicated in grey. For morphotypes having some specimens included in a different MOTU, the collecting sites of those individuals are also reported. Vertical lines indicate MOTUs and GROUPS as identified in the analysis. Numbers at each node represent BPP and NJ, MP and ML bootstrap values. Only significant values (BPP > 0.95 and bootstrap >80%) are reported.

Figure 4.

Bayesian phylogenetic tree for the native Mediterranean Cerithium specimens based on the combined COI and 12S mitochondrial gene (cmtDNA) dataset and rooted with C. scabridum. Names of terminal taxa correspond to the scored haplotypes (listed in Table 3). Morphotypes doubtfully (?) included indicated in grey. For morphotypes having some specimens included in a different MOTU, the collecting sites of those individuals are also reported. Vertical lines indicate MOTUs and GROUPS as identified in the analysis. Numbers at each node represent BPP and NJ, MP and ML bootstrap values. Only significant values (BPP > 0.95 and bootstrap >80%) are reported.

Species delimitation

The results of species delimitation methods applied to the cmtDNA dataset conformed to the phylogenetic analyses. Thus, the 20 recursive steps in the ABGD analysis resulted in six different sequence partitions (Table 5; Fig. 5), ranging from one to 13 PSH. The best correspondence between PSH and the MOTUs identified in the phylogenetic analyses was found at Pmax = 0.005456 (partition 8), where all the PSH corresponded to the main lineages. Partitions 9, 10 and 11 corresponded only partially to the MOTUs, with one group pooling MOTU-A and MOTU-B together. The first seven (0.001 < Pmax < 0.004) and the last five partitions (0.014 < Pmax < 0.037) were not considered due to the excessive splitting or lumping of identified morphotypes, respectively. For the COI dataset (Supplementary Material Table S1, Fig. S9), the number of groups defined by the recursive partition in the ABGD analysis was 12 with 0.001 < Pmax < 0.004; 8 with prior of 0.005; 7 with prior of 0.006; 6 with priors of 0.008 and 0.011; 4 with 0.010 < Pmax < 0.023; 3 with priors of 0.029 and 0.037; 1 with prior of 0.048. The primary partition was stable on the range of prior values with eight groups, only four of which corresponded to MOTUs identified in the phylogenetic analyses (MOTU-3, MOTU-4, MOTU-5 and MOTU-10).

Table 5.

Results of ABGD analysis for cmtDNA dataset.

Partition number Groups P-value 
1–4 13 0.001000–0.002069 
5–7 12 0.002637–0.004281 
0.005456 
9–11 0.006952–0.011288 
12–15 0.014384–0.029764 
16 0.037927 
Partition number Groups P-value 
1–4 13 0.001000–0.002069 
5–7 12 0.002637–0.004281 
0.005456 
9–11 0.006952–0.011288 
12–15 0.014384–0.029764 
16 0.037927 

Analysis was performed using the K2P model to calculate pairwise distances, 20 recursive steps, X = 0.5, Pmin = 0.001, Pmax = 0.1, Nb = 20 (where P = prior maximum intraspecific divergence).

Figure 5.

Histogram representing the frequency distribution of pairwise distance values among combined COI-12S sequences (cmtDNA) for the Mediterranean Cerithium specimens, with the evident barcode gap found in the dataset corresponding to the maximum value of intraspecific differences calculated by the ABGD analysis. Pairwise distance values were calculated using the K2P model.

Figure 5.

Histogram representing the frequency distribution of pairwise distance values among combined COI-12S sequences (cmtDNA) for the Mediterranean Cerithium specimens, with the evident barcode gap found in the dataset corresponding to the maximum value of intraspecific differences calculated by the ABGD analysis. Pairwise distance values were calculated using the K2P model.

The results of the analysis using the SDP are shown in Table 6. All MOTUs comprised nonrandom monophyletic groups, with the highest Intra Dist values reported for clades A, F and G, indicating that the divergence within these groups was high relative to the divergence from their closest MOTU. All the Rosenberg's PAB values were significant (P < 0.001), indicating that all MOTUs are putative species, while all Rodrigo's P(RD) values were lower than 0.05 except for MOTU-F, thus suggesting six species. The GSI calculated for each identified MOTU were values of one and highly significant (P < 0.001 after 10,000 permutation tests), thus rejecting the null hypothesis that the lineages are of mixed ancestry. For the GMYC analysis, no significant difference was observed when single and multiple speciation events were applied (χ2 = 2.185, P = 0.624), so only results of the single-threshold model are reported (Fig. 6). The ML of the null model of uniform branching across the tree (i.e. sequences are drawn from a single species) was significantly lower than the ML of the GMYC model (null: 161.5979, GMYC: 164.1311, likelihood-ratio: 5.066, P = 0.007). The threshold indicating the single speciation coalescent transition occurred at a value of 0.117 Ma, recovering 12 ML clusters, five of which exactly match the lineages identified in the phylogenetic analyses (MOTU-A, B, C, D and E) and the remaining seven including one or two individuals of MOTU-F and MOTU-G.

Table 6.

Summary statistics for each MOTU (i.e. putative species) reported by SDP analysis for the cmtDNA dataset.

MOTU Closest species Intra Inter Intra/Inter P ID (strict) P ID (liberal) Rosenberg's PAB Rodrigo's P(RD) 
0.224 0.480 0.47 0.82 (0.75, 0.89) 0.95 (0.91, 0.99) 4.3e−5 0.044 
0.068 0.480 0.14 0.70 (0.52, 0.87) 0.93 (0.78, 1.0) 2.2e−5 0.048 
0.101 0.600 0.17 0.76 (0.61, 0.90) 0.94 (0.83, 1.0) 6.5e−6 0.045 
0.109 0.752 0.14 0.88 (0.77, 0.99) 0.96 (0.90, 1.0) 3.6e−4 0.045 
0.120 0.752 0.16 0.76 (0.62, 0.90) 0.95 (0.84, 1.0) 3.6e−4 0.046 
0.425 0.786 0.54 0.57 (0.44, 0.70) 0.85 (0.75, 0.95) 2.98e−03 0.048 
0.270 0.786 0.34 0.56 (0.38, 0.74) 0.81 (0.67, 0.96) 2.98e−03 0.047 
MOTU Closest species Intra Inter Intra/Inter P ID (strict) P ID (liberal) Rosenberg's PAB Rodrigo's P(RD) 
0.224 0.480 0.47 0.82 (0.75, 0.89) 0.95 (0.91, 0.99) 4.3e−5 0.044 
0.068 0.480 0.14 0.70 (0.52, 0.87) 0.93 (0.78, 1.0) 2.2e−5 0.048 
0.101 0.600 0.17 0.76 (0.61, 0.90) 0.94 (0.83, 1.0) 6.5e−6 0.045 
0.109 0.752 0.14 0.88 (0.77, 0.99) 0.96 (0.90, 1.0) 3.6e−4 0.045 
0.120 0.752 0.16 0.76 (0.62, 0.90) 0.95 (0.84, 1.0) 3.6e−4 0.046 
0.425 0.786 0.54 0.57 (0.44, 0.70) 0.85 (0.75, 0.95) 2.98e−03 0.048 
0.270 0.786 0.34 0.56 (0.38, 0.74) 0.81 (0.67, 0.96) 2.98e−03 0.047 

Abbreviations: Intra/inter, ratio of intra (genetic differentiation among members of a putative species) to inter (genetic differentiation between the members of a putative species and the members of the closest putative species); P ID (strict), mean probability of correctly identifying an unknown number of a given clade using the criterion that it must fall within, but not sister, to the species clade in a tree; P ID (liberal), mean probability, with 95% confidence interval, of correctly identifying an unknown specimen placed on a tree, with the criterion that it falls sister to or within a monophyletic species clade; Rosenberg's PAB, probability of reciprocal monophyly under a random coalescent model; Rodrigo's P(RD), probability that a clade has the observed degree of distinctiveness due to random coalescent processes.

Figure 6.

Likelihood profile of the single-threshold GMYC model in the native Mediterranean Cerithium detected from combined COI–12S sequences (cmtDNA). The solid vertical line indicates the threshold for the maximum likelihood GMYC model transition, indicating the shift from phylogenetic Yule processes to coalescent population processes. Grey branches indicate putative species. MOTUs as identified in the other analyses indicated on the right.

Figure 6.

Likelihood profile of the single-threshold GMYC model in the native Mediterranean Cerithium detected from combined COI–12S sequences (cmtDNA). The solid vertical line indicates the threshold for the maximum likelihood GMYC model transition, indicating the shift from phylogenetic Yule processes to coalescent population processes. Grey branches indicate putative species. MOTUs as identified in the other analyses indicated on the right.

DISCUSSION

Molecular markers and species delimitation methods

Although species richness is often estimated by using the COI gene alone (‘DNA barcoding’; Hebert et al., 2003), there are some doubts concerning species identification based on single-gene approaches (e.g. Wiemers & Fiedler, 2007; Lohse, 2009). In a single-gene framework, incomplete lineage-sorting or introgression can result in underestimation of species diversity (Funk & Omland, 2003) and phylogeographic breaks in the distribution of haplotypes may result in the erroneous recognition of populations as separate species, even when there are no barriers to gene flow. Inability of the barcoding region to recover species has been observed in several studies (e.g. Sauer & Hausdorf, 2012), showing that it can lack sufficient information to discriminate species. An integrated approach based on more genes therefore appears necessary for accurate species delimitation, especially in the case of closely related species. Yet, it is still not clear which genes are directly involved in the speciation process, or how many DNA regions are necessary for species delimitation. Although molecular analyses have been based largely on mitochondrial genes, it is increasingly acknowledged that species inference using mtDNA should be corroborated with information from nuclear loci (O'Meara, 2010; Ross et al., 2010), but conflicts can arise between mitochondrial and nuclear data because of different gene evolutionary histories (e.g. Bichain et al., 2007). In this work, two mitochondrial genes (COI and 12S) were employed to explore the taxonomic status of several nominal Cerithium species from the central Mediterranean Sea. All analyses based on the combined dataset gave the same results, indicating the potential suitability of COI and 12S as a barcode for the discrimination of Cerithium species. However, the resulting preliminary species hypotheses must be tested with additional data and the addition of nuclear data is a logical and necessary subsequent step to this study, to formulate robust and stable taxonomic interpretations.

Several different approaches for delimiting species based on molecular data have recently been developed (e.g. Wiens, 2007), but it is difficult to know which method should be used for any given dataset, or how many analyses are necessary to obtain robust species hypotheses. All methods are based on different assumptions and show advantages and disadvantages. In this study, we used four species-delimitation methods to explore the taxonomic status of the nominal Cerithium species, one based on genetic distances (ABGD) and three based on tree topology (SDP, GSI and GMYC). The independence of ABGD analysis from tree topology is considered an advantage, because the method does not rely on properties of internal nodes of the species tree. As such, while not delimiting species under the PSC, ABGD is considered an appropriate approach for identifying putative species (Puillandre et al., 2012) and was used herein to obtain a set of species hypotheses to be further explored with other methods (Kekkonen & Hebert, 2014). On the other hand, SDP, GSI and GMYC rely on the recognition of monophyletic groups and hence these methods allow delimiting species under the PSC. The results of ABGD, SDP and GSI were concordant, thus indicating their potential utility in defining Cerithium species. The GMYC recovered 14 putative species, most of which represent singletons, but the model is known to overestimate the number of MOTUs compared to other methods (e.g. Talavera, Dincă & Vila, 2013), especially when groups involved in the coalescent process have been incompletely sampled (e.g. Papadopoulou et al., 2009). Consequently, it is possible that our incomplete sampling has resulted in the artificial inflation of putative species. Given this result, the following discussion is based on the phylogenetic analyses and the ABGD, SDP and GSI analyses only.

Morphological vs molecular Cerithium diversity

Phylogenetic analyses and the methods employed in this study for delimiting species based on the combined COI and 12S genes show that patterns of morphological and genetic variation are not completely in agreement. Our DNA sequence data allowed the identification of seven MOTUs, four of which coincided with recognized morphotypes (MOTU-A = C. vulgatum, MOTU-B = C. alucastrum, MOTU-C = C. protractum and MOTU-E = C. lividulum). The other three MOTUs contained mixtures of haplotypes from different morphotypes (MOTU-D and MOTU-F), or formed a distinct clade, but with geographic structure (MOTU-G).

MOTU-D is represented by large individuals including all those identified as C. repandum and the western (O1/cmt15 from Sardinia) and southern (24/cmt16 from Apulia) individuals of C. vulgatum morphotypes. The apparent polyphyly of C. vulgatum is due to the lack of clearcut discriminating characters to separate it from C. repandum. Indeed, the nominal C. vulgatum and C. repandum are extremely similar (Monterosato, 1884; Cecalupo, Buzzurro & Mariani, 2008) and are currently considered synonyms, so differences in habitat are often used to differentiate the two species, with C. repandum living in very shallow water on fine sand or mud, and in lagoons, presumably only in its type locality in the Gulf of Gabès, and with C. vulgatum occupying deeper Posidonia beds, algal bottoms or bare rocks in the open sea throughout the Mediterranean. Therefore, because of their geographic origin and their habitat, it is likely that Tunisian individuals represent true C. repandum, and that specimens from Oristano (O1) and Brindisi (24) were misidentified as C. vulgatum and are actually conspecific with C. repandum. Consequently, MOTU-D probably represents the large C. repandum which, if a valid species, would not be endemic of the Gulf of Gabès.

Similarly, MOTU-F represents a clade of small individuals with a mixture of morphotypes that correspond to the general teleoconch morphology of C. renovatum, and a ‘morphological’ C. lividulum from Tunisia. Although the morphological identification of individuals of MOTU-F was more certain than specimens of MOTU-D, the mixture of morphotypes in the former indicates that the distinctive keel thought to be diagnostic of C. renovatum is not always present, making identification difficult. Further investigation including a high number of C. lividulum and C. renovatum is required to understand better the relationships between the two species and the characters that may reliably separate them.

A clear geographic pattern emerged in MOTU-G, including all C. renovatum from Crete, representing a probable cryptic species. This is not an unexpected finding in a group such as Cerithium with species that tend to be morphologically difficult to define (Miglietta, Faucci & Santini, 2011), but additional specimens are necessary to test this hypothesis. Similar C. renovatum-like individuals have been found on Zakynthos, western Greece, probably representing yet another distinct species endemic to the Eastern basin (S. Gofas, personal communication), which also requires further testing.

Utility of morphological characters in taxonomy of Cerithium

Reflecting well the problem between intra- and interspecific variation among Mediterranean Cerithium, only partial concordance was found between traditional morphological and sequence data. This suggests that morphological characters traditionally used in their classification are not always reliable. In general, the match among MOTU-A, MOTU-B, MOTU-C and the morphotypes C. vulgatum, C. alucastrum, C. protractum, respectively, supports well the value of the morphological characters used for their diagnosis. On the contrary, the mixture of C. vulgatum and C. repandum morphotypes in MOTU-D underlines the problematic morphological identification of C. repandum based on teleoconch morphology and the importance of the protoconch in differentiating it from C. vulgatum. The distribution of C. lividulum and C. renovatum morphotypes over three clades (MOTU-E, MOTU-F and MOTU-G), highlights the problematic morphological identification of C. renovatum based on the teleoconch, with characters such as keels and knobs proving unreliable, and the protoconch being the most important character for distinguishing it from C. lividulum. This finding supports Gofas et al. (2004), who pointed out that only some characters of the adult shell of C. lividulum and C. renovatum are useful to distinguish between the two in the case of sympatric populations, and Garilli & Galletti (2006), who stated that C. lividulum and C. renovatum are both highly polymorphic, with the larval shell the only character able to distinguish them and to recognize atypical morphotypes.

The molecular data in this study suggest that, based on traditional taxonomy, morphological variants may have been inappropriately lumped into a single species or subdivided into separate species. As a result, Mediterranean Cerithium diversity is probably currently underestimated. These findings underline the importance of an interdisciplinary approach for constructing robust taxonomic hypotheses at the species level in taxa such as Mediterranean Cerithium that exhibit a lack of concordance between morphological species and MOTUs.

How many native species of Cerithium live in the Mediterranean Sea and what are the relationships among them?

Establishing the number of native Cerithium species living in the Mediterranean Sea is a difficult task, which will require a great deal of comparative morphological and molecular data and adequate geographic sampling of the area. The data obtained in this study of Cerithium from the central basin should provide a starting point for defining species in future revisionary works. Here, we have taken a conservative approach, considering valid only the so-called ‘stable’ species (Padial & De la Riva, 2010), i.e. those supported by morphological and genetic lines of evidence that do not conflict, with the caveat that the protoconch as a key discriminating character was often unavailable to us.

Using this approach, three large species can be unambiguously distinguished among the examined specimens: C. vulgatum, C. alucastrum and C. protractum. The fact that the morphological and molecular distinctiveness of these species are maintained in sympatry (i.e. Pag Island, Chioggia) is particularly significant. This finding resolves the controversy surrounding the distinctiveness of the three. Previously, deeper water C. alucastrum and shallow-water C. vulgatum sometimes have been considered as ecotypes, while C. protractum has sometimes been regarded as a littoral morph of C. alucastrum or an elongate form of C. vulgatum (e.g. Gofas, Moreno & Salas, 2011). Our results demonstrate clearly that they are three distinct species, with C. alucastrum and C. vulgatum distinguished by the presence of axial costae and the absence of typical spiny nodules in the latter (Monterosato, 1910). The data also corroborate Gofas et al. (2011), who tentatively treated C. protractum as valid pending further testing with molecular data. The large C. repandum is also apparently valid, but individuals O1 and 24 indicate the presence of morphological variants in regions outside its assumed range in the Gulf of Gabès. Based on the present results, the small C. lividulum and C. renovatum may represent more than two species, but this hypothesis requires examination of additional material for thorough testing.

Perhaps unsurprisingly, C. vulgatum (MOTU-A), C. alucastrum (MOTU-B) and C. protractum (MOTU-C), all large species, are strongly linked (GROUP1, Fig. 4), with a closer association between C. vulgatum and C. alucastrum. Similarly, small individuals in MOTU-F (C. renovatum+C. lividulum from Tunisia) and MOTU-G (C. renovatum from Crete), were supported as strongly linked (GROUP 3, Fig. 4). Unexpectedly, the large Cerithium included in MOTU-D is not supported as closely linked with other large species, and small C. lividulum (MOTU-E) is not supported as closely linked to other small species. This pattern is difficult to explain, but is similar to that recovered by Boisselier-Dubayle & Gofas (1999), showing that a population of the small C. renovatum clusters with the local large C. vulgatum and not with C. lividulum (the other recognized small species). However, no specific geographic patterns were observed between MOTU-E and MOTU-D, as they include haplotypes from both northern and southern localities. Assuming that size is a useful phylogenetic character, these findings suggest that COI and 12S, while useful for exploring Cerithium species diversity, may not be suitable for recovering relationships among them. Only the analysis of additional loci, as well as more comprehensive geographic sampling, can resolve these questions.

SUPPLEMENTARY MATERIAL

Supplementary Material is available at Journal of Molluscan Studies online.

ACKNOWLEDGEMENTS

We are grateful to those who provided samples for this study: Paolo Giulio Albano, Gianni Spada, Francesca Calandriello, Michela Kuan, Paolo Russo and Marco Taviani. We also would like to acknowledge the staff of the Molecular Systematics Laboratory, Department of Invertebrates, Royal Belgian Institute of Natural Sciences, Brussels, where the genetic analyses started. In particular many thanks to Thierry Backeljau, Karin Breugelmans and Vanya Prévot. We also thank Marco Passamonti, Department of Biological, Geological and Environmental Sciences, University of Bologna, Italy, who provided helpful suggestions for data interpretation. Paolo Giulio Albano, Institut für Paläontologie, Universität Wien, Austria, significantly improved the manuscript with constructive and insightful comments. The authors want to thank two reviewers, Marco Oliverio and Serge Gofas, for their valuable comments and suggestions on an earlier draft. Many thanks also to Associate Editor Ellen E. Strong and Editor David G. Reid for their revision of the manuscript. This work was supported by the Canziani Fund, Department of Biological, Geological and Environmental Sciences, University of Bologna. Partial financial support was also provided by the Marco Polo Grant from the University of Bologna Francesca Evangelisti.

REFERENCES

BARASH
A.
,
ZENZIPER
Z.
1980
.
Egg masses of Mollusca from the Mediterranean waters of Israel and notes on reproduction of the freshwater species Theodoxus jordani and Melanoides tubercolata
.
Veliger
 ,
22
:
299
317
.
BICHAIN
J.M.
,
GAUBER
P.
,
SAMADI
S.
,
BOISSELIER-DUBAYLE
M.C.
2007
.
A gleam in the dark: phylogenetic species delimitation in the confusing spring-snail genus Bythinella Moquin-Tandon, 1856 (Gastropoda: Rissooidea: Amnicolidae)
.
Molecular Phylogenetics and Evolution
 ,
45
:
927
941
.
BLAXTER
M.L.
2004
.
The promise of a DNA taxonomy
.
Philosophical Transactions of the Royal Society of London B
 ,
359
:
669
679
.
BOISSELIER-DUBAYLE
M.C.
,
GOFAS
S.
1999
.
Genetic relationships between marine and marginal marine populations of Cerithium species from the Mediterranean Sea
.
Marine Biology
 ,
135
:
671
682
.
BROCCHI
G.B.
1814
.
Conchiologia fossile subappenninica con osservazioni geologiche sugli Appennini e sul suolo adiacente
 .
Silvestri
,
Milano
.
BUCQUOY
E.
,
DAUTZEMBERG
P.
,
DOLFUS
G.
1884
.
Les mollusques marins du Roussillon, Tome I, Gasteropodes
 .
Baillière
,
Paris
.
CECALUPO
A.
2006
.
Elenco della famiglia Cerithiidae de Férussac, 1819 (Prosobranchia). Nomenclatura delle specie conosciute o poco note e relativa revisione sistematica. Attuali
.
Quaderni della Civica Stazione Idrobiologica di Milano
 ,
27
:
1
373
.
CECALUPO
A.
,
BUZZURRO
G.
,
MARIANI
M.
2008
.
Contributo alla conoscenza della malacofauna del Golfo di Gabès (Tunisia)
.
Quaderni della Civica Stazione Idrobiologica di Milano
 ,
31
:
175
.
CUMMINGS
M.P.
,
NEEL
M.C.
,
SHAW
K.L.
2008
.
A genealogical approach to quantifying lineage divergence
.
Evolution
 ,
62
:
2411
2422
.
DAYRAT
B.
2005
.
Towards integrative taxonomy
.
Biological Journal of the Linnean Society
 ,
85
:
407
415
.
DE QUEIROZ
K.
1998
.
The general lineage concept of species, species criteria, and the process of speciation
. In:
Endless forms: species and speciation
  (
Howard
D.
,
Berlocher
S.
, eds), pp.
57
75
.
Oxford University Press
,
Oxford
.
DE QUEIROZ
K.
2005a
.
Different species problems and their resolution
.
Bioassays
 ,
27
:
1263
1269
.
DE QUEIROZ
K.
2005b
.
A unified concept of species and its consequences for the future of taxonomy
.
Proceedings of the California Academy of Sciences
 ,
56
:
196
215
.
DE QUEIROZ
K.
2007
.
Species concepts and species delimitation
.
Systematic Biology
 ,
56
:
879
886
.
ERIXON
P.
,
SVENNBLAD
B.
,
BRITTON
T.
,
OXELMAN
B.
2003
.
Reliability of Bayesian posterior probabilities and bootstrap frequencies in phylogenetics
.
Systematic Biology
 ,
52
:
665
673
.
FOLMER
O.
,
BLACK
M.
,
HOEH
W.
,
LUTZ
R.
,
VRIJENHOEK
R.
1994
.
DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates
.
Molecular Marine Biology and Biotechnology
 ,
3
:
294
299
.
FONTANETO
D.
,
HERNIOU
E.A.
,
BOSCHETTI
C.
,
CAPRIOLI
M.
,
MELONE
G.
,
RICCI
C.
,
BARRACLOUG
T.G.
2007
.
Independently evolving species in asexual bdelloid rotifers
.
PLoS Biology
 ,
5
:
914
921
.
FUJITA
M.K.
,
LEACHÉ
A.D.
,
BURBRINK
F.T.
,
MCGUIRE
J.A.
,
MORITZ
C.
2012
.
Coalescent-based species delimitation in an integrative taxonomy
.
Trends in Ecology and Evolution
 ,
27
:
480
488
.
FUNK
D.J.
,
OMLAND
K.E.
2003
.
Species-level paraphyly and polyphyly: frequency, causes and consequences, with insights from animal mitochondrial DNA
.
Annual Review of Ecology, Evolution and Systematics
 ,
34
:
397
423
.
GARILLI
V.
,
GALLETTI
L.
2006
.
Taxonomical characters for distinguishing Cerithium lividulum Risso, 1826, and C. renovatum Monterosato, 1884 (Gastropoda, Caenogastropoda, Cerithiidae)
.
Basteria
 ,
70
:
109
122
.
GIANNUZZI-SAVELLI
R.
,
PUSATERI
F.
,
PALMERI
A.
,
EBREO
C.
1997
.
Atlante delle conchiglie marine del Mediterraneo
 .
La Conchiglia
,
Roma
.
GOFAS
S.
2015
.
Cerithium protractum. In: MolluscaBase (2015). Accessed through: World Register of Marine Species
at .
GOFAS
S.
,
GARILLI
V.
,
BOISSELIER-DUBAYLE
M.C.
2004
.
Nomenclature of the smaller Mediterranean Cerithium species
.
Bollettino Malacologico
 ,
39
:
95
104
.
GOFAS
S.
,
MORENO
D.
,
SALAS
C.
2011
.
Moluscos marinos de Andalucía
 .
Servicio de Publicaciones e Intercambio Científico, Universidad de Málaga
,
Málaga
.
HALL
T.A.
1999
.
BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT
.
Nucleic Acids Symposium Series
 ,
41
:
95
98
.
HEBERT
P.D.N.
,
PENTON
E.H.
,
BURNS
J.M.
,
JANZEN
D.H.
,
HALLWACHS
W.
2003
.
Ten species in one: DNA barcoding reveals cryptic species in the neotropical skipper butterfly Astraptes fulgerator
.
Proceedings of the National Academy of Sciences of the USA
 ,
101
:
14812
14817
.
HOUBRICK
R.S.
1974
.
The genus Cerithium in the Western Atlantic
.
Johnsonia
 ,
5
:
33
84
.
HOUBRICK
R.S.
1976
.
The family Cerithiidae in the Indo-Pacific. Part 1: the genera Rhinoclavis, Pseudovertagus and Clavocerithium
.
Monograph of Marine Mollusca
 ,
1
:
1
130
.
HOUBRICK
R.S.
1985
.
Genus Clypeomorus Jousseaume (Cerithiidae: Prosobranchia)
.
Smithsonian Contributions to Zoology
 ,
403
:
1
131
.
HOUBRICK
R.S.
1992
.
Monograph of the genus Cerithium Bruguière in the Indo-Pacific (Cerithiidae: Prosobranchia)
.
Smithsonian Contributions to Zoology
 ,
510
:
1
211
.
HURVICH
C.M.
,
TSAI
C.L.
1989
.
Regression and time series model selection in small samples
.
Biometrika
 ,
76
:
297
307
.
HUSON
D.H.
,
SCORNAVACCA
C.
2012
.
Dendroscope 3: an interactive tool for rooted phylogenetic trees and networks
.
Systematic Biology
 ,
61
:
1061
1067
.
KAWAHARA
A.Y.
,
OHSHIMA
I.
,
KAWAKITA
A.
,
REGIER
J.C.
,
MITTER
C.
,
CUMMINGS
M.P.
,
DAVIS
D.R.
,
WAGNER
D.L.
,
DE PRINIS
J.
,
LOPEZ-VAAMONDE
C.
2011
.
Increased gene sampling strengthens support for higher-level groups within leaf-mining moths and relatives (Lepidoptera: Gracillariidae)
.
BMC Evolutionary Biology
 ,
11
:
182
.
KEKKONEN
M.
,
HEBERT
P.D.N.
2014
.
DNA barcode-based delineation of putative species: efficient start for taxonomic workflows
.
Molecular Ecology Resources
 ,
14
:
706
707
.
KOCHER
T.D.
,
THOMAS
W.K.
,
MEYER
A.
,
EDWARDS
S.V.
,
PÄÄBO
S.
,
VILLABLANCA
F.X.
,
WILSON
A.C.
1989
.
Dynamics of mitochondrial DNA evolution in animals: amplification and sequencing with conserved primers
.
Proceedings of the National Academy of Sciences of the USA
 ,
86
:
6169
6200
.
LIBRADO
P.
,
ROZAS
J.
2009
.
DnaSP v5.10: a software for comprehensive analysis of DNA polymorphism data
.
Bioinformatics
 ,
25
:
1451
1451
.
LOHSE
K.
2009
.
Can mtDNA barcodes be used to delimit species? A response to Pons et al. (2006)
.
Systematic Biology
 ,
58
:
439
442
.
MALLET
J.
2001
.
Species, concepts of
. In:
Encyclopedia of biodiversity
  (
Levin
S.A.
, ed.), pp.
427
440
.
Academic Press
,
New York
.
MASTERS
B.C.
,
FAN
V.
,
ROSS
H.A.
2011
.
Species delimitation—a Geneious plugin for the exploration of species boundaries
.
Molecular Ecology Resources
 ,
11
:
154
157
.
MIGLIETTA
M.P.
,
FAUCCI
A.
,
SANTINI
F.
2011
.
Speciation in the sea: overview of the symposium and discussion of future directions
.
Integrative and Comparative Biology
 ,
51
:
449
455
.
MIURA
O.
,
TORCHIN
M.E.
,
KURIS
A.M.
,
HECHINGER
R.F.
,
CHIBA
S.
2006
.
Introduced cryptic species of parasites exhibit different invasion pathways
.
Proceedings of the National Academy of Sciences of the USA
 ,
103
:
19818
19823
.
MONAGHAN
M.T.
,
WILD
R.
,
ELLIOT
M.
,
FUJISAWA
T.
,
BALKE
M.
,
INWARD
D.L.G.
,
LEES
D.C.
,
RANAIVOSOLO
R.
,
EGGLETON
P.
,
BARRACLOUGH
T.
,
VOGLER
A.P.
2009
.
Accelerated species inventory on Madagascar using coalescent-based models of species delineation
.
Systematic Biology
 ,
58
:
298
311
.
MONTEROSATO
T.
1884
.
Nomenclatura generica e specifica di alcune conchiglie mediterranee
 .
Stabilimento Tipografico Virzì
,
Palermo
.
MONTEROSATO
T.
1910
.
Nota su taluni generi e specie della famiglia Cerithiidae
.
Giornale di Scienze Naturali ed Economiche di Palermo
 ,
28
:
65
76
.
MONTSERRAT
R.
1994
.
Spawn morphology and features of the protoconch of Cerithium alucaster (Brocchi, 1814) and Euspira fusca (Blainville, 1825) (Gastropoda: Prosobranchia)
.
Scientia Marina
 ,
58
:
185
190
.
NORDSIECK
F.
1974
.
Il genere Thericium Monterosato nei mari d'Europa
.
Conchiglia
 ,
59
:
3
12
.
O'MEARA
B.C.
2010
.
New heuristic methods for joint species delimitation and species tree inference
.
Systematic Biology
 ,
59
:
59
73
.
OLIVERIO
M.
,
GIANNUZZI-SAVELLI
R.
,
GIOVINE
F.
,
MANGANELLI
G.
,
PUSATERI
F.
,
SABELLI
B.
,
SPADA
G.
2008
.
Gastropoda Prosobranchia
. In:
Checklist della flora e della fauna dei mari italiani
  (
Relini
G.
, ed.), pp.
247
278
.
Biologia Marina Mediterranea
,
Genova
.
PADIAL
J.M.
,
DE LA RIVA
I.
2010
.
A response to recent proposals for integrative taxonomy
.
Biological Journal of the Linnean Society
 ,
101
:
747
756
.
PALOMBI
A.
1939
.
Uova e larve di Cerithium rupestre Risso (Gastropoda: Prosobranchia)
.
Bollettino di Zoologia
 ,
10
:
209
212
.
PAPADOPOULOU
A.
,
MONAGHAN
M.T.
,
BARRACLOUGH
T.G.
,
VOGLER
A.P.
2009
.
Sampling error does not invalidate the Yule-coalescent model for species delimitation. A response to Lohse (2009)
.
Systematic Biology
 ,
58
:
442
444
.
PAVLINOV
I.Y.
2013
.
The species problem, why again
. In:
The species problem—ongoing issues
  (
Pavlinov
I.Y.
, ed.), pp.
3
37
.
InTech
,
Rijeka
.
PONS
J.
,
BARRACLOUGH
T.G.
,
GOMEZ-ZURITA
J.
,
CARDOSO
A.
,
DURAN
D.P.
,
HAZELL
S.
,
KAMOUN
S.
,
SUMLIN
W.D.
,
VOGLER
A.P.
2006
.
Sequence-based species delimitation for the DNA taxonomy of undescribed insects
.
Systematic Biology
 ,
55
:
595
609
.
PUILLANDRE
N.
,
LAMBERT
A.
,
BROUILLET
S.
,
ACHAZ
G.
2012
.
ABGD, automatic barcode gap discovery for primary species delimitation
.
Molecular Ecology
 ,
21
:
1864
1877
.
RODRIGO
A.
,
BERTELS
F.
,
HELED
J.
,
NODER
R.
,
SHEARMAN
H.
,
TSAI
P.
2008
.
The perils of plenty: what are we going to do with all these genes?
Philosophical Transactions of the Royal Society of London B
 ,
363
:
3893
3902
.
RONQUIST
F.
,
TESLENKO
M.
,
VAN DER MARK
P.
,
AYRES
D.
,
DARLING
A.
,
HÖFNA
S.
,
LARGET
B.
,
LIU
L.
,
SUCHARD
M.A.
,
HUELSENBECK
J.P.
2012
.
MrBayes 3.2: efficient Bayesian phylogenetic inference and model choices across a large model space
.
Systematic Biology
 ,
61
:
539
542
.
ROSENBERG
N.A.
2007
.
Statistical tests for taxonomic distinctiveness from observations of monophyly
.
Evolution
 ,
61
:
317
323
.
ROSS
K.G.
,
GOTZEK
D.
,
ASCUNCE
M.S.
,
SHOEMAKER
D.D.
2010
.
Species delimitation: a case study in a problematic ant taxon
.
Systematic Biology
 ,
59
:
162
184
.
SAUER
J.
,
HAUSDORF
B.
2012
.
A comparison of DNA-based methods for delimiting species in a Cretan land snail radiation reveals shortcomings of exclusively molecular taxonomy
.
Cladistics
 ,
28
:
300
316
.
SCHLICK-STEINER
B.C.
,
STEINER
F.M.
,
SEIFERT
B.
,
STAUFFER
C.
,
CHRISTIAN
E.
,
CROZIER
R.H.
2010
.
Integrative taxonomy: a multisource approach to exploring biodiversity
.
Annual Review of Entomology
 ,
55
:
421
438
.
SCHWARZ
G.
1978
.
Estimating the dimension of a model
.
Annals of Statistics
 ,
6
:
461
464
.
SHIMAYAMA
T.
,
HIMENO
H.
,
SASUGA
J.
,
YOKOBORI
S.
,
UEDA
T.
,
WATANABE
K.
1990
.
The genetic code of a squid mitochondrial gene
.
Nucleic Acids Symposium Series
 ,
22
:
77
78
.
SITES
J.W.
,
MARSHALL
J.C.
2003
.
Delimiting species: a renaissance issue in systematic biology
.
Trends in Ecology and Evolution
 ,
18
:
462
470
.
SUGIURA
N.
1978
.
Further analysis of the data by Akaike's information criterion and the finite corrections
.
Communications in Statistics—Theory and Methods
 ,
7
:
13
26
.
SWOFFORD
D.L.
2002
.
PAUP*. Phylogenetic analysis using parsimony (and other methods). Version 4.0b10
.
Sinauer Associates
,
Sunderlands, MA
.
TALAVERA
G.
,
DINCĂ
V.
,
VILA
R.
2013
.
Factors affecting species delimitations with the GMYC model: insights from a butterfly survey
.
Methods in Ecology and Evolution
 ,
4
:
1101
1110
.
TAMURA
K.
,
PETERSON
D.
,
PETERSON
N.
,
STECHER
G.
,
NEI
M.
,
KUMAR
S.
2011
.
MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods
.
Molecular Biology and Evolution
 ,
28
:
2731
2739
.
THOMSPON
J.D.
,
HIGGINS
D.G.
,
GIBSON
T.J.
1994
.
CLUSTALW: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, positions-specific gap penalties and weight matrix choice
.
Nucleic Acids Research
 ,
22
:
4673
4680
.
WIEMERS
M.
,
FIEDLER
K.
2007
.
Does the DNA barcoding gap exist?—a case study in blue butterflies (Lepidoptera: Lycaenidae)
.
Frontiers in Zoology
 ,
4
:
8
.
WIENS
J.J.
2007
.
Species delimitation: new approaches for discovering diversity
.
Systematic Biology
 ,
56
:
875
878
.
WINNEPENNINCKX
B.T.
,
BACKELJAU
T.
,
DE WACHTER
R.
1993
.
Complete small ribosomal subunit RNA sequence of the chiton Acanthopleura japonica (Lischke, 1873) (Mollusca, Polyplacophora)
.
Nucleic Acids Research
 ,
21
:
1670
.
XIA
X.
,
XIE
Z.
2001
.
DAMBE: software package for data analysis in molecular biology and evolution
.
Journal of Heredity
 ,
92
:
371
373
.
XIA
X.
,
XIE
Z.
,
SALEMI
M.
,
CHEN
L.
,
WANG
Y.
2003
.
An index of substitution saturation and its application
.
Molecular Phylogenetics and Evolution
 ,
26
:
1
7
.